Dear Dominique,
Thank you for the permission to use the code. I hope we can set up
all of Mathilde's code to soon work with relax. I now have some older
code from Paul for 2-site exchange with complex conjugate matrices
(http://thread.gmane.org/gmane.science.nmr.relax.devel/4132) very
close to be up and running within relax. Hence adding the different
models that Mathilde has sent should be quite easy. I already have
parts of her code in relax.
For the copyright assignment, I will ask Mathilde and Paul to check
each of files, one-by-one, and to say which they have contributed to.
For a copyright statement, legally both the persons name and the years
when changes were made must be present. Also, legally, without a
formal copyright transfer instrument, only those who wrote the code
are allowed to have their names in the copyright notice. But we can
list all names where it counts - within the documentation. Also, for
those who wish, as authors for a short letter I am considering for all
of the relaxation dispersion contributions of the relevant relax
authors and contributors (in Bioinformatics), essentially just an
announcement. I'll discuss the documentation later with Paul and
Mathilde.
The code that is incorporated into relax, once complete, should always
remain in a working state. I plan on having tests added to the relax
test suite to make sure that the code is functional and will remain
that way for the length of the relax project. As this is an open
source project with many contributors, this does not depend on me
always present to maintain the project.
I hope that you will be able to make use of advanced tools present
within relax (starting from peak list reading and ending in multiple
data visualisations), together with the dispersion code from your
group, in the future. One major advantage, for example, is the
optimisation in relax. The current code uses the Levenberg-Marquardt
algorithm but, as the gradients and Jacobians cannot be calculated,
this is performed numerically within Scipy
(http://docs.scipy.org/doc/scipy/reference/generated/scipy.optimize.leastsq.html).
A numerical technique with a subnested numerical approximation is not
ideal for speed. I also know that that Scipy optimisation code is, or
at least was, horribly buggy as I mentioned recently at
http://article.gmane.org/gmane.science.nmr.relax.user/1445:
"You might also be interested in learning about the minfx project that
relax uses for optimisation (https://gna.org/projects/minfx/). This
originated as a relax package as the scipy optimisers all contained
fatal bugs back in 2003 (I'm not sure they have been fixed as the
original developers were MIA even back then and I think have never
returned). But it was spun out into its own software distribution."
For optimisation using the much more advanced minfx package
(http://gna.org/projects/minfx/), relax uses the simplex or
Nelder-Mead algorithm together with the log-barrier constraint
algorithm for the dispersion analysis. The first does not use the
gradient or Jacobian significantly speeding up the optimisation, the
second can also massively speed up optimisation in many situations.
In the future, if someone has the need and interest in implementing
it, relax's multi-processor infrastructure could also be used to speed
up calculations.
I hope that the code from your group together with the code already
present in relax will result in the most useful and simplest to use
software for relaxation dispersion. Once all the code is
incorporated, relax will also be the most comprehensive package for
relaxation dispersion analyses as both the analytical and numerical
solutions will be supported (and can be directly compared). And it is
incredibly easy for users who use the GUI. An official relax release
with all this code will be made when everyone is ready and happy with
the state of the analysis.
Sincerely,
Edward
P. S. Incorporating code or even having someone become a relax
developer is a relatively easy process. It only requires a
demonstration of adhering to the relax coding style and of high code
quality. If you find relax useful and you have ideas for advancing
the dispersion analysis, you are most welcome to consider this as an
option.
On 15 July 2013 11:13, Dominique Marion (IBS) <Dominique.Marion@xxxxxx> wrote:
Dear Edward,
Mathilde Lescanne wrote the code for the CPMG fit while she was a master
student under my supervision. We started from the code that Paul Schanda
wrote few years ago. In this context, I grant you permission to include
this piece of code into the Relax software as distributed under the GNU
General Public Licence and agree to have my name among the author list.
Mathilde is moving next autumn to a dutch lab for her PhD and I will try to
maintain the code with her help in the future.
Best regards,
Dominique
_______________________________________________
relax (http://www.nmr-relax.com)
This is the relax-devel mailing list
relax-devel@xxxxxxx
To unsubscribe from this list, get a password
reminder, or change your subscription options,
visit the list information page at
https://mail.gna.org/listinfo/relax-devel
Re: Copyright notice for the mpower() function